Author + information
- Received April 22, 2015
- Revision received June 26, 2015
- Accepted July 24, 2015
- Published online October 13, 2015.
- Kavitha N. Pundi, MD∗,
- Jonathan N. Johnson, MD∗,†,
- Joseph A. Dearani, MD‡,
- Krishna N. Pundi, BS§,
- Zhuo Li, BS‖,
- Cynthia A. Hinck, RN, BSN∗,
- Sonja H. Dahl, RN, DNP∗,
- Bryan C. Cannon, MD∗,†,
- Patrick W. O’Leary, MD∗,†,
- David J. Driscoll, MD∗ and
- Frank Cetta, MD∗,†∗ ()
- ∗Division of Pediatric Cardiology, Mayo Clinic, Rochester, Minnesota
- †Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- ‡Division of Cardiovascular Surgery, Mayo Clinic, Rochester, Minnesota
- §Mayo Clinic College of Medicine, Mayo Clinic, Rochester, Minnesota
- ‖Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota
- ↵∗Reprint requests and correspondence:
Dr. Frank Cetta, Division of Pediatric Cardiology, Mayo Clinic, Gonda 6335, 200 First Street SW, Rochester, Minnesota 55905.
Background There are limited long-term, single-cohort, follow-up studies available about patients after the Fontan operation.
Objectives This study sought to determine the long-term outcome of all patients who had a Fontan operation at the Mayo Clinic.
Methods Records of all patients who had a modified Fontan operation between 1973 and 2012 were reviewed. A follow-up questionnaire was mailed to all patients alive at the time of the study.
Results Overall, 10-, 20-, and 30-year survival for 1,052 patients was 74%, 61%, and 43%, respectively. Factors associated with decreased overall or late survival in multivariate analysis included pre-operative diuretic use, longer cardiopulmonary bypass time, operation prior to 1991, atrioventricular valve (AVV) replacement at the time of Fontan operation, elevated post-bypass Fontan (>20 mm Hg) or left atrial (>13 mm Hg) pressures, prolonged chest tube drainage (>21 days), post-operative ventricular arrhythmias, renal insufficiency, and development of protein-losing enteropathy (PLE). Pre-operative and intraoperative sinus rhythm were associated with improved survival. Long-term survival was similar for patients regardless of ventricular morphology. The most common reoperations were pacemaker insertion/revision in 212 patients (20%), Fontan revision/conversion in 117 patients (11%), and AVV repair/replacement in 66 patients (5%). Clinically significant late atrial or ventricular arrhythmias occurred in 468 patients (44%). Ninety-five patients (9%) developed PLE, and 5-, 10-, and 20-year survival after diagnosis of PLE was 50%, 35%, and 19%, respectively.
Conclusions As the surgical techniques for the Fontan operation have changed over the last 40 years, survival has improved. However, development of PLE and arrhythmias and the need for reoperation during long-term follow-up pose significant management challenges.
In 1971, Fontan and Baudet described a surgical technique for successful palliation of patients with tricuspid atresia (1,2). Subsequently, this technique has been applied to treat most forms of functional single ventricles (3–7). Theoretically, the Fontan operation separates the systemic and pulmonary venous returns to ameliorate the disadvantages of long-term hypoxemia, reduce thromboembolic events, preserve ventricular function, and prolong survival for patients with single-ventricle physiology. Although some of these beliefs have been fulfilled, a number of adverse results of the Fontan procedure have been recognized, including premature death, ventricular failure, thromboembolic disease, arrhythmia, liver disease, and protein-losing enteropathy (PLE) (8–14). In this study, we sought to determine long-term outcomes for all patients who had a Fontan operation at our institution.
In this institutional review board–approved, single-center, retrospective study, we reviewed the records of all patients (n = 1,052) who had their initial Fontan operation at the Mayo Clinic between October 1973 and June 2012 and who provided informed consent to participate in the study. Information regarding demographic, anatomic, pre-operative, operative, and post-operative variables, and Mayo follow-up was abstracted into a secure electronic database. Many of the patients had their follow-up care at other institutions. Any available correspondence regarding echocardiogram, cardiac catheterization, electrocardiogram, Holter/event monitor, laboratory tests, liver imaging, or surgical/procedural data was entered into the patient database.
A medical questionnaire was mailed to all patients not known to be dead at the initiation of the study. Nonresponders received second and third questionnaires, and if these were not returned or completed, an attempt was made to contact the patients by telephone. Quality-of-life surveys from patients rated their current health status on a scale of 1 to 4 (1 = excellent, 2 = good, 3 = fair, and 4 = poor). Data regarding death were updated using correspondence from physicians and patients/family members, chart review, and the Social Security Death Index. For the remaining patients, data curves were censured using the date of last available follow-up or date of transplant.
PLE was diagnosed based on documentation of enteric loss of alpha-1-antitrypsin or the presence of low serum total protein/albumin in addition to persistent or intermittent edema. Patients were excluded from subsequent PLE analysis if the timing of diagnosis was not known. Clinically significant arrhythmia was defined as the need for antiarrhythmic drug therapy (excluding digoxin), pacemaker placement, or electrical/pharmacological cardioversion. Patients with arrhythmias prior to the Fontan procedure were excluded from any analysis of post-operative arrhythmias. Cirrhosis was diagnosed based on liver biopsy/autopsy or characteristic findings on computed tomography, cardiac magnetic resonance imaging, magnetic resonance elastography, or ultrasound in conjunction with clinical diagnosis by a gastroenterologist. Patients with isolated liver function or ultrasound abnormalities were not considered to have proven cirrhosis.
Variables used in the Cox regression analysis were initially analyzed as continuous variables, and then discrete cutoffs were selected based on the hazard ratios. The cutoffs for discrete variables used in univariate/multivariate analyses were defined as follows: pre-operative pulmonary artery pressure (PAP) (>17 mm Hg), pre-operative systemic ventricular end-diastolic pressure (>12 mm Hg), pre-operative pulmonary arteriolar resistance (>3 U × m2), post-bypass left atrial (LA) pressure (>13 mm Hg), post-bypass Fontan pressure (>20 mm Hg), and prolonged chest tube duration (≥21 days).
Separate analyses were performed using either the date of the Fontan operation as time 0 (“overall survival”) or 30 days after the operation as time 0 (“late survival”). All deaths, regardless of cause, after the Fontan operation were considered in the survival analysis. Descriptive statistics for categorical variables were reported as frequency and percentage, and continuous variables were reported as mean ± SD or median (range) as appropriate. Time to PLE was compared between Fontan type groups using analysis of variance. Kaplan-Meier curves were derived to calculate 10-, 20-, and 30-year survival statistics. Cox regression models were used to determine univariate and multivariate predictors of survival and other long-term outcomes. The multivariable model considered significant univariate variables (p < 0.05) with model selection using the stepwise method. All statistical tests were 2-sided with the alpha level set at 0.05 for statistical significance. SAS version 9.3 (SAS Institute, Inc., Cary, North Carolina) was used for the analysis. The set of variables evaluated for association of survival and long-term outcomes and results of univariate Cox regression analyses are listed in Online Tables 1 to 4.
Between 1973 and 2012, 1,052 patients had an initial Fontan operation at the Mayo Clinic (Table 1). At last follow-up, 426 patients (40%) were known to be dead. Of the 626 patients known to be alive, transplant-free survival was verified in 427 patients (68%) with follow-up information within 5 years of the study termination date. The mean age at initial Fontan procedure was 9.4 ± 7.5 years (median: 7 years; range: 7 months to 53 years). Mean follow-up after the Fontan operation was 15.3 ± 9.3 years (median: 15.1 years; range: 34 days to 37 years). The oldest survivor after the Fontan operation was 67 years of age (Fontan at age 39 years). Of the 723 follow-up questionnaires mailed out, 305 (42%) were returned.
Overall, 10-, 20-, and 30-year survival after the Fontan operation was 74%, 61%, and 43%, respectively (Figure 1A). Overall survival by surgical era, type of Fontan procedure, and pre-operative anatomy are depicted in Figures 1B, 2, and 3, respectively. Kaplan-Meier estimates of overall survival are listed in Table 2, and factors associated with decreased overall survival in multivariate analysis are in Table 3. When only pre-operative variables were considered, factors associated with decreased overall survival in multivariate analysis included operation prior to 1991, use of pre-operative diuretics, asplenia, and lack of pre-operative sinus rhythm. Fenestration was associated with improved overall survival only on univariate analysis. However, only 86 patients (8%) underwent fenestration at the time of their initial Fontan procedure (most of which were performed after 1995). Ventricular morphology did not impact overall survival.
The 10-, 20-, and 30-year survival for the 932 patients who were alive 30 days after the Fontan operation was 83%, 68%, and 48% respectively. Kaplan-Meier estimates of late survival are listed in Table 4. Factors associated with decreased late survival in multivariate analysis are listed in Table 5. When only pre-operative variables were considered, factors associated with decreased late survival included operation prior to 1991, use of pre-operative diuretics or amiodarone, interrupted inferior vena cava (heterotaxy), and lack of pre-operative sinus rhythm.
Early reoperations and complications
A total of 177 of 1,052 patients (17%) had at least 1 early reoperation after their Fontan operation. The 3 most common indications for reoperation were bleeding (n = 74 [42%]), wound debridement (n = 38 [21%]), and placement of a permanent pacemaker (n = 35 [20%]). There were 21 patients who had early Fontan takedown (<30 days after Fontan operation). Of these, 12 patients (57%) died during long-term follow-up (9 [43%] died within 30 days of takedown). The incidence of atrial arrhythmias (n = 224) and ventricular arrhythmias (n = 86) prior to hospital discharge was 21% and 8%, respectively.
Of the 426 patients known to be deceased, cause of death was known for 281 patients (66%). A total of 234 of 281 deaths (83%) were from a primary cardiac cause; however, the cause of death was multifactorial in many patients. Contributing factors in the reported cause of death for 281 patients included respiratory failure (n = 101 [36%]), renal insufficiency (n = 85 [30%]), sudden death or arrhythmia (n = 52 [19%]), bleeding complications or cardiac tamponade (n = 50 [18%]), disseminated intravascular coagulopathy/sepsis (n = 48 [17%]), PLE (n = 31 [11%]), and hepatic insufficiency (n = 29 [10%]). Four patients died from malignancies: 2 with hepatocellular carcinoma, 1 with metastatic cervical cancer, and 1 with multiple myeloma.
The 10-, 20-, and 30-year freedom from death or cardiac transplant was 73%, 59%, and 40%, respectively. Factors associated with decreased survival or transplant in multivariate analysis included elevated pre-operative PAP (>17 mm Hg), asplenia, use of pre-operative diuretics, operation prior to 1991, longer bypass time, absence of intraoperative sinus rhythm, and elevated post-bypass Fontan (>20 mm Hg) or LA (>13 mm Hg) pressures (Table 6). Thirty-eight patients in the cohort had heart transplants (mean age at transplant 24.0 ± 11.8 years), with 24 of 38 (63%) known to be alive post-transplant (mean follow-up 10.4 ± 6.9 years).
Among the 932 patients who survived at least 30 days after the initial Fontan operation, the 10-, 20-, and 30-year freedom from death or reoperation was 69%, 50%, and 33%, respectively. Factors associated with decreased survival or reoperation in multivariate analysis included asplenia, use of pre-operative diuretics, elevated pre-operative PAP (>17 mm Hg), atriopulmonary type of Fontan connection, atrioventricular valve (AVV) replacement at the time of Fontan procedure, intraoperative rhythm other than sinus, longer bypass time, and elevated post-bypass Fontan pressure (>20 mm Hg) (Table 7). A prior bidirectional Glenn procedure was associated with improved survival and freedom from reoperation.
Most patients had their long-term follow-up at other medical institutions. The most common late reoperations were pacemaker insertion/revision (n = 212 [23%]), Fontan revision/conversion (n = 117 [13%]), and AVV repair/replacement (n = 66 [7%]). Implantable cardioverter-defibrillators were placed in 14 patients (2%) during long-term follow-up. At least 71 of 117 patients (61%) undergoing Fontan conversion/revision had a prior diagnosis of arrhythmia at a mean duration of 6.3 ± 5.6 years prior to their operation. Of these 117 patients, 10 (9%) died within 30 days of their Fontan conversion/revision. During long-term follow-up, 37 of 117 patients (32%) died at a mean duration of 5.2 ± 5.7 years (median 2.7 years) after reoperation. There were 17 patients who had late Fontan takedown; of these, 12 patients (71%) died during long-term follow-up (7 [41%] within 30 days of Fontan takedown).
Freedom from PLE at 10, 20, and 30 years after the Fontan procedure was 92%, 89%, and 83%, respectively. A total of 95 patients were diagnosed with PLE, of whom 88 patients had an available date of diagnosis. For these 88 patients, mean age at Fontan was 11.2 ± 8 years and mean interval from Fontan to diagnosis of PLE was 8.1 ± 7.9 years (median: 5 years; range: 2 months to 32 years). Fifty-one of 88 patients with PLE (58%) had a date of diagnosis prior to 1995. Fenestration at the time of Fontan procedure was performed in 4 patients; 8 patients had late fenestration as an attempted therapy for PLE. There was no significant difference in mean duration to onset of PLE by type of Fontan operation. Factors associated with development of PLE in multivariate analysis are listed in Table 8.
Overall mortality in the PLE cohort was 72% (63 of 88) during 7.0 ± 7.4 years of follow-up. Cause of death was known for 36 of 88 patients: chronic heart failure in 13, sepsis/multiorgan failure in another 13, and other PLE-related complications in the remaining 10. Survival at 5, 10, and 20 years after PLE diagnosis was 50%, 35%, and 19%, respectively (Central Illustration).
There were 996 of 1,052 patients (95%) who did not have a known pre-operative arrhythmia; they were included in arrhythmia analyses. For these patients, the overall freedom from post-operative arrhythmias at 10, 20, and 30 years after the Fontan operation was 71%, 42%, and 24%, respectively. New arrhythmias were diagnosed more than 30 days after the Fontan operation in 412 of 996 patients (41%), with some patients having multiple arrhythmias. Late arrhythmias in these 412 patients included atrial flutter (n = 304 [74%]), atrial fibrillation (n = 161 [39%]), atrial tachycardia (n = 108 [26%]), re-entrant supraventricular tachycardia (n = 37 [9%]), and ventricular tachycardia (n = 40 [10%]).
A total of 195 patients had available liver imaging or biopsy/pathology data and of these, 40 patients (21%) were diagnosed with cirrhosis at a mean duration of 23.3 ± 6.3 years from the Fontan procedure; their average age at the time of Fontan operation was 9.6 ± 9.2 years. Five patients were diagnosed with hepatocellular carcinoma at a mean duration of 20 ± 2.9 years after the Fontan operation (2 patients died during long-term follow-up).
We present the largest long-term follow-up study from a single institution for patients after the Fontan operation.
Mortality, transplant, and reoperation
As surgical techniques and management of patients after the Fontan operation have improved over the last 40 years, a number of investigators have reported improvement in early and intermediate-term survival (9,11,15). In our cohort, early mortality after Fontan operation was 13% in the surgical era prior to 1991 but decreased to 7.8% from 1991 to 2000 and 6.9% in the era following 2000. This increase in patient survival has been attributed to multiple factors, including better patient selection and improvement in operative techniques and post-operative patient management.
Indeed, studies from the Mayo Clinic—based on our early cohort—contributed to the development of more appropriate selection criteria (8,10). Compared with a similar cohort of patients having the Fontan operation at Boston Children’s Hospital (median age 7.9 years), there was a comparable high incidence of early and late mortality after the initial Fontan operation (12). However, unlike the Boston study, in which there was significant attrition in late follow-up data, our 20-year data were based on 35% of the patients in our cohort. Many of the factors associated with long-term mortality or transplant in the Boston study (PLE, diuretic therapy, higher Fontan pressure) were consistent with our study’s results.
Recent studies from the Children’s Hospital of Philadelphia (15), the Australia/New Zealand registry (11,16), and the University of Alabama at Birmingham (9) noted that early and late survival after Fontan operation was excellent, with no increase in late mortality. However, these results are based on a more recent surgical era than patients in our series, and the Fontan connections were primarily lateral tunnel or extracardiac Fontan connections involving younger patients with a higher proportion of fenestrations. Additionally, the survival data at 20 years and beyond in these studies were based on a very small number of patients. The long-term survival of patients with lateral tunnel Fontan connections was poor in our study, likely due to several reasons: 1) many of the earlier “lateral tunnel” operations were performed in patients with heterotaxy syndromes, for which intra-atrial conduits were used; and 2) others incorporated a variable amount of native atrial tissue into the surgical repair, making them susceptible to some of the same long-term complications as operations in patients with atriopulmonary connections. Additionally, ventricular morphology did not influence overall survival in this study, presumably because a majority of the right ventricular morphology patients in this study did not have hypoplastic left heart syndrome (therefore, there was no need for a Norwood operation).
The timing of Fontan conversion, or listing for cardiac transplant, in a patient with a long-standing atriopulmonary Fontan connection is not always straightforward. In this study, we noted that the <30-day mortality with Fontan conversion/revision was 9%. Studies from multiple institutions (17–19) have identified risk factors for poor outcome with Fontan conversion, such as PLE, systemic right ventricular/intermediate ventricular morphology, significant AVV regurgitation, older age (>27 years), elevated systemic ventricular end-diastolic pressure, renal dysfunction, and cirrhosis. However, either concomitant arrhythmia operation or concomitant surgery for Fontan obstruction is associated with better overall outcome. The Boston Children’s group evaluated the outcomes of failing Fontan patients listed for transplant and noted decreased survival in patients with preserved ventricular function compared with those with impaired ventricular function (20). This was attributed to the challenges of determining the ideal timing of cardiac transplant in patients with preserved ventricular function because it has been shown that long-term survival is reasonable in patients with PLE despite increased early post-operative mortality after transplant (21,22). However, in the older atriopulmonary Fontan patients who do not clearly meet criteria for a low-risk Fontan conversion or listing for transplant, management decisions need to be made on an individual basis after careful consideration of associated risks and benefits.
In other large single-center studies, approximately 2% of patients required late cardiac transplant, and there has been limited long-term follow-up (11,12). In our cohort, 38 patients (3.6%) had heart transplants (mean age at transplant 24.0 ± 11.8 years), and 63% of these patients were known to be alive at last follow-up.
Abnormalities related to Fontan physiology
The overall incidence of PLE in this study was 9%, which is similar to the 5% to 15% incidence quoted in other large studies (23–25). Freedom from developing PLE at 10, 20, and 30 years post-Fontan procedure was 92%, 89%, and 83%, respectively. Survival at 5, 10, and 20 years after diagnosis of PLE was 50%, 35%, and 19%, respectively. As noted in previous studies, elevated pre-operative mean PAP and post-bypass LA pressure were associated with an increased risk of developing PLE (23). A contemporary cohort of patients with PLE followed at our institution had much better survival (25); however, this study included patients who did not have their Fontan operations at our institution. Hence, it is not a cohort study, and one cannot directly compare those survival statistics with those of a cohort study such as ours. Nonetheless, the substantial mortality in our cohort would suggest that aggressive management of PLE and earlier consideration of heart transplant might be important in these patients, especially if the PLE is refractory to medical management.
Overall, 44% of patients had a diagnosis of a new clinical arrhythmia during long-term follow-up. Freedom from arrhythmia at 10, 20, and 30 years after the Fontan operation was 71%, 42%, and 24%, respectively. A majority of these patients had atrial flutter or fibrillation, and a smaller proportion had re-entrant supraventricular tachycardia, atrial tachycardia, or ventricular tachycardia.
Increased attention is being paid to the development of cirrhosis and hepatocellular carcinoma in patients who have had the Fontan operation (26–32). Presumably, this results at least in part due to elevated hepatic venous pressure, but understanding of the pathophysiology of liver dysfunction after Fontan procedure is evolving. There is likely a strong referral bias accounting for the high incidence of cirrhosis in our patient cohort. However, it still highlights the need for more aggressive evaluation and management of liver disease in these patients. The role of cardiac transplant (with or without concomitant liver transplant) continues to be debated in these patients (21,33–40).
Quality of life
Eighty percent of the patients after the Fontan operation rated their current health as excellent, and a similar percentage of patients thought that their physical status was improved. There is likely a selection bias in these data because the patients who chose to complete these surveys may be healthier than their counterparts and had their Fontan surgery during a later era. However, this is consistent with previously reported data suggesting that patients tend to perceive themselves as having a higher functional status compared with control populations (41).
One limitation of our study is that it was a single-center cohort from a large tertiary referral center over an almost 4-decade time frame. Therefore, in all the late follow-up outcomes, the data curves were censured using the date of last available follow-up on each patient. Some of these results may not be directly translatable to a younger cohort of patients in a recent surgical era. Liver disease may be underreported in this cohort because imaging studies were not routinely performed in patients operated on in the early surgical era. However, as more patients with congenital heart disease live into adulthood, these data are relevant in the medical management of patients with atriopulmonary or lateral tunnel Fontan operations, especially those who have survived 20 to 30 years after their initial operation.
With better patient selection and evolution of surgical techniques and medical management of patients after the Fontan operation, there has been improvement in survival over the last 40 years. The overall 10-, 20-, and 30-year survival in our study from the time of operation was 74%, 61%, and 43%, respectively. However, the development of PLE, ventricular failure, cirrhosis, arrhythmias, and the need for reoperation during long-term follow-up pose significant management challenges.
COMPETENCY IN MEDICAL KNOWLEDGE: Over the last 40 years, surgical techniques used in the Fontan operation have evolved and survival rates have improved, but post-operative protein-losing enteropathy and arrhythmias and the need for reoperation during long-term follow-up remain considerable challenges.
TRANSLATIONAL OUTLOOK: Further long-term studies are needed to examine the occurrence of late complications in patients who have undergone atriopulmonary Fontan procedures.
The authors thank Drs. Dwight McGoon, Gordon Danielson, Francisco Puga, and Harold Burkhart, who performed many of the operations in this study, and Matt Cetta for assistance with data abstraction.
For supplemental tables, please see the online version of this article.
Dr. Cannon has served as a consultant to Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrioventricular valve
- left atrial
- pulmonary artery pressure
- protein-losing enteropathy
- systemic ventricular end-diastolic pressure
- Received April 22, 2015.
- Revision received June 26, 2015.
- Accepted July 24, 2015.
- American College of Cardiology Foundation
- Puga F.J.
- Cetta F.,
- Feldt R.H.,
- O’Leary P.W.,
- et al.
- Driscoll D.J.,
- Offord K.P.,
- Feldt R.H.,
- et al.
- d’Udekem Y.,
- Iyengar A.J.,
- Galati J.C.,
- et al.
- Khairy P.,
- Fernandes S.M.,
- Mayer J.E. Jr..,
- et al.
- Bartz P.J.,
- Driscoll D.J.,
- Dearani J.A.,
- et al.
- Deal B.J.,
- Jacobs M.L.
- Rogers L.S.,
- Glatz A.C.,
- Ravishankar C.,
- et al.
- Iyengar A.J.,
- Winlaw D.S.,
- Galati J.C.,
- et al.
- van Son J.A.,
- Mohr F.W.,
- Hambsch J.,
- et al.
- Gambetta K.,
- Backer C.,
- Deal B.,
- et al.
- John A.S.,
- Johnson J.A.,
- Khan M.,
- et al.
- Samsky M.D.,
- Patel C.B.,
- DeWald T.A.,
- et al.
- Michielon G.,
- van Melle J.P.,
- Wolff D.,
- et al.
- Bernstein D.,
- Naftel D.,
- Chin C.,
- et al.